💥Thought Shrapnel newsletter #442

Coming at you from Substack...

Doug Belshaw

Oct 1, 2023

👋 Hello!

How are you doing? Well, I hope.

Situation report: I’m sitting down to compose this at 14:53 on Saturday afternoon. My son is getting his stuff ready for his basketball game, while my daughter’s football team won 16-3 this morning. It’s a mild autumnal day in Northumberland.

Check out Weeknote 39/2023 for exciting news about my new car and other solo waffle.

DID YOU KNOW that with Substack you can not only like/favourite this, but also comment on it. Imagine that! Audience feedback FTW.

💥 Best of Thought Shrapnel

As usual, I published 16 posts on Thought Shrapnel this week. Of those, here are three to which I'd like to draw to your attention.

Felled tree at Sycamore Gap along Hadrian's Wall

Sycamore Stump

There is, or rather was, a tree that symbolised the North East of England. Standing at a dip in the ground along Hadrian’s Wall called ‘Sycamore Gap’, it’s a tree I’ve visited many times with friends and family. Last year, when I walked the wall in 72 hours, it was a familiar touchstone.

Now the iconic 200 year-old tree, which featured in the film Robin Hood: Prince of Thieves, is gone. Felled by a 16 year-old in an act of wanton vandalism. On a World Heritage Site. Some people just want to watch the world burn.

It didn’t take long for someone to rename the place on Google Maps where the tree used to stand to ‘Sycamore Stump’. Hopefully they will build some kind of memorial to it. I do think it’s difficult for someone not from the region to understand just how important things like this are to one’s identity.

A 16-year-old boy has been arrested in northern England in connection with what authorities described as the “deliberate” felling of a famous tree that had stood for nearly 200 years next to the Roman landmark Hadrian’s Wall.
[…]
Photographs from the scene on Thursday showed the tree was cut down near the base of its trunk, with the rest of it lying on its side.
Northumbria Police said the teen was arrested on suspicion of causing criminal damage. He was in police custody and assisting officers with their inquiries.
[…]
“This is an incredibly sad day,” police Superintendent Kevin Waring said. “The tree was iconic to the North East and enjoyed by so many who live in or who have visited this region.”

Source: ‘Incredibly sad day’: Teen arrested in England after felling ancient tree | Al Jazeera

Image: Oli Scarff / Agence France-Presse (taken from NYT article)

Robot teacher with chalk next to chalkboard

‘Personalisation’ is something that humans do

Audrey Watters, formerly the ‘Cassandra’ of edtech, is now writing about health, nutrition, and fitness technologies at Second Breakfast. It’s great, I’m a paid subscriber.

In this article, she looks at the overlap between her former and current fields, comparing and contrasting coaches and educators with algorithms. While I don’t share her loathing of ChatGPT, as an educator and a parent I’d definitely agree that motivation and attention is something to which a human is (currently) best suited.

How well does a teacher or trainer or coach know how you feel, how well you performed, or what you should do or learn next? How well does an app know how you feel, how well you performed, or what you should do next? Digital apps insist that, thanks to the data they collect, they can make better, more precise recommendations than humans ever can — dismissing what humans do as “one size fits all.” Yet it’s impossible to scrutinize their algorithmic decision-making. Ideally, at least, you can always ask your coach, “Why the hell am I doing bulgarian split squats?! These suck.” And she will tell you precisely why. (Love you, Coach KB.)
And then (ideally) she’ll say, “If you don’t want to do them, you don’t have to.” And (ideally), she’ll ask you what’s going on. Maybe you feel like shit that day. Maybe you don’t have time. Maybe they hurt your hamstrings. Maybe you’d like to hear some options — other exercises you can do instead. Maybe you’d like to know why she prescribed this exercise in the first place — “it’s a unilateral exercises, and as a runner,” she says, “we want to work on single-leg strength, with a focus on your glute medius and adductors because I’ve noticed, by watching your barbell squats, that those areas are your weak spots.” This is how things get “personalized” — not by some massive data extraction and analysis, but by humans asking each other questions and then tailoring our responses and recommendations accordingly. Teachers and coaches do this every. goddamn. day. Sure, there’s a training template or a textbook that one is supposed to follow; but good teachers and coaches check in, and they switch things up when they’re not really working.
[…]
If we privilege these algorithms, we’re not only adopting their lousy recommendations; we’re undermining the expertise of professionals in the field. And we’re not only undermining the expertise of professionals in the field, we’re undermining our own ability to think and learn and understand our own bodies. We’re undermining our own expertise about ourselves. (ChatGPT is such a bad bad bad idea.)

Source: Teacher/Coach as Algorithm | Second Breakfast

AI generated images with subliminal messages

You’ve probably seen some of these already. Someone discovered that if you use the generator for QR codes but feed it something different, it can create words from images.

There are lots more examples at knowyourmeme and you can try creating your own using KREA.

ControlNet uses the AI image-generating tool Stable Diffusion, and one of its initial uses was generating fancy QR codes using the code as an input image. That idea was then taken further, with some users developing a workflow that lets them specify any image or text as a black-and-white mask that implants itself into the generated image—kind of like an automated, generative version of the masking tool in Photoshop.
“What happened there was that this user discovered that if they used the QR Code ControlNet but instead of feeding it a QR code, they fed it some other black-and-white patterns, they could create nice optical illusions,” said Passos. “You can now send a conditioning image and the model blends in a pattern that satisfies that while still making a coherent image at the same time.”

Source: AI-Generated ‘Subliminal Messages’ Are Going Viral. Here’s What’s Really Going On | VICE

✍️ The rest of Thought Shrapnel

Until next week!

Thought Shrapnel Weekly is published by Dr. Doug Belshaw. You can connect with him by replying to this email, or via LinkedIn or the Fediverse.

Many thanks to Bryan Mathers of Visual Thinkery for the Thought Shrapnel logo.

All product names, logos, and brands are property of their respective owners and are used in this newsletter are for identification purposes only.

Currently reading

Troubled Blood by Robert Galbraith
The Mushroom at the End of the World by Anna Lowenhaupt Tsing.

I also read (and re-read) a book called Secret Tradecraft of Elite Advisors by David C. Baker this week, which was pretty great.

🤘 Super-secret link to reward those who scroll to the bottom of newsletters!